4 # :title: rbot utilities provider
6 # Author:: Tom Gilbert <tom@linuxbrit.co.uk>
7 # Author:: Giuseppe "Oblomov" Bilotta <giuseppe.bilotta@gmail.com>
9 # TODO some of these Utils should be rewritten as extensions to the approriate
10 # standard Ruby classes and accordingly be moved to extends.rb
15 # Try to load htmlentities, fall back to an HTML escape table.
17 require 'htmlentities'
110 AFTER_PAR_PATH = /^(?:div|span)$/
111 AFTER_PAR_EX = /^(?:td|tr|tbody|table)$/
112 AFTER_PAR_CLASS = /body|message|text/i
118 # Some regular expressions to manage HTML data
121 TITLE_REGEX = /<\s*?title\s*?>(.+?)<\s*?\/title\s*?>/im
124 HX_REGEX = /<h(\d)(?:\s+[^>]*)?>(.*?)<\/h\1>/im
126 PAR_REGEX = /<p(?:\s+[^>]*)?>.*?<\/?(?:p|div|html|body|table|td|tr)(?:\s+[^>]*)?>/im
128 # Some blogging and forum platforms use spans or divs with a 'body' or 'message' or 'text' in their class
129 # to mark actual text
130 AFTER_PAR1_REGEX = /<\w+\s+[^>]*(?:body|message|text)[^>]*>.*?<\/?(?:p|div|html|body|table|td|tr)(?:\s+[^>]*)?>/im
132 # At worst, we can try stuff which is comprised between two <br>
133 AFTER_PAR2_REGEX = /<br(?:\s+[^>]*)?\/?>.*?<\/?(?:br|p|div|html|body|table|td|tr)(?:\s+[^>]*)?\/?>/im
140 # Miscellaneous useful functions
142 @@bot = nil unless defined? @@bot
143 @@safe_save_dir = nil unless defined?(@@safe_save_dir)
150 # Set up some Utils routines which depend on the associated bot.
152 debug "initializing utils"
154 @@safe_save_dir = @@bot.path('safe_save')
161 SEC_PER_HR = SEC_PER_MIN * 60
163 SEC_PER_DAY = SEC_PER_HR * 24
165 SEC_PER_WK = SEC_PER_DAY * 7
166 # Seconds per (30-day) month
167 SEC_PER_MNTH = SEC_PER_DAY * 30
168 # Second per (non-leap) year
169 SEC_PER_YR = SEC_PER_DAY * 365
171 # Auxiliary method needed by Utils.secs_to_string
172 def Utils.secs_to_string_case(array, var, string, plural)
175 array << "1 #{string}"
177 array << "#{var} #{plural}"
181 # Turn a number of seconds into a human readable string, e.g
182 # 2 days, 3 hours, 18 minutes and 10 seconds
183 def Utils.secs_to_string(secs)
185 years, secs = secs.divmod SEC_PER_YR
186 secs_to_string_case(ret, years, _("year"), _("years")) if years > 0
187 months, secs = secs.divmod SEC_PER_MNTH
188 secs_to_string_case(ret, months, _("month"), _("months")) if months > 0
189 days, secs = secs.divmod SEC_PER_DAY
190 secs_to_string_case(ret, days, _("day"), _("days")) if days > 0
191 hours, secs = secs.divmod SEC_PER_HR
192 secs_to_string_case(ret, hours, _("hour"), _("hours")) if hours > 0
193 mins, secs = secs.divmod SEC_PER_MIN
194 secs_to_string_case(ret, mins, _("minute"), _("minutes")) if mins > 0
196 secs_to_string_case(ret, secs, _("second"), _("seconds")) if secs > 0 or ret.empty?
199 raise "Empty ret array!"
203 return [ret[0, ret.length-1].join(", ") , ret[-1]].join(_(" and "))
207 # Turn a number of seconds into a hours:minutes:seconds e.g.
208 # 3:18:10 or 5'12" or 7s
210 def Utils.secs_to_short(seconds)
211 secs = seconds.to_i # make sure it's an integer
212 mins, secs = secs.divmod 60
213 hours, mins = mins.divmod 60
215 return ("%s:%s:%s" % [hours, mins, secs])
217 return ("%s'%s\"" % [mins, secs])
219 return ("%ss" % [secs])
223 # Returns human readable time.
227 # :start_date, sets the time to measure against, defaults to now
228 # :date_format, used with <tt>to_formatted_s<tt>, default to :default
229 def Utils.timeago(time, options = {})
230 start_date = options.delete(:start_date) || Time.new
231 date_format = options.delete(:date_format) || "%x"
232 delta = (start_date - time).round
236 distance = Utils.age_string(delta)
238 _("%{d} from now") % {:d => distance}
240 _("%{d} ago") % {:d => distance}
245 # Converts age in seconds to "nn units". Inspired by previous attempts
246 # but also gitweb's age_string() sub
247 def Utils.age_string(secs)
250 Utils.age_string(-secs)
251 when secs > 2*SEC_PER_YR
252 _("%{m} years") % { :m => secs/SEC_PER_YR }
253 when secs > 2*SEC_PER_MNTH
254 _("%{m} months") % { :m => secs/SEC_PER_MNTH }
255 when secs > 2*SEC_PER_WK
256 _("%{m} weeks") % { :m => secs/SEC_PER_WK }
257 when secs > 2*SEC_PER_DAY
258 _("%{m} days") % { :m => secs/SEC_PER_DAY }
259 when secs > 2*SEC_PER_HR
260 _("%{m} hours") % { :m => secs/SEC_PER_HR }
261 when (20*SEC_PER_MIN..40*SEC_PER_MIN).include?(secs)
263 when (50*SEC_PER_MIN..70*SEC_PER_MIN).include?(secs)
264 # _("about one hour")
266 when (80*SEC_PER_MIN..100*SEC_PER_MIN).include?(secs)
267 _("an hour and a half")
268 when secs > 2*SEC_PER_MIN
269 _("%{m} minutes") % { :m => secs/SEC_PER_MIN }
271 _("%{m} seconds") % { :m => secs }
277 # Execute an external program, returning a String obtained by redirecting
278 # the program's standards errors and output
280 # TODO: find a way to expose some common errors (e.g. Errno::NOENT)
282 def Utils.safe_exec(command, *args)
283 output = IO.popen("-") { |p|
285 break p.readlines.join("\n")
288 $stderr.reopen($stdout)
290 rescue Exception => e
291 puts "exception #{e.pretty_inspect} trying to run #{command}"
294 puts "exec of #{command} failed"
298 raise "safe execution of #{command} returned #{$?}" unless $?.success?
302 # Try executing an external program, returning true if the run was successful
303 # and false otherwise
304 def Utils.try_exec(command, *args)
308 $stderr.reopen($stdout)
310 rescue Exception => e
322 # Safely (atomically) save to _file_, by passing a tempfile to the block
323 # and then moving the tempfile to its final location when done.
325 # call-seq: Utils.safe_save(file, &block)
327 def Utils.safe_save(file)
328 raise 'No safe save directory defined!' if @@safe_save_dir.nil?
329 basename = File.basename(file)
330 temp = Tempfile.new(basename,@@safe_save_dir)
332 yield temp if block_given?
334 File.rename(temp.path, file)
338 # Decode HTML entities in the String _str_, using HTMLEntities if the
339 # package was found, or UNESCAPE_TABLE otherwise.
341 def Utils.decode_html_entities(str)
342 if defined? ::HTMLEntities
343 return HTMLEntities.decode_entities(str)
345 str.gsub(/(&(.+?);)/) {
347 # remove the 0-paddng from unicode integers
348 if symbol =~ /^#(\d+)$/
349 symbol = $1.to_i.to_s
352 # output the symbol's irc-translated character, or a * if it's unknown
353 UNESCAPE_TABLE[symbol] || (symbol.match(/^\d+$/) ? [symbol.to_i].pack("U") : '*')
358 # Try to grab and IRCify the first HTML par (<p> tag) in the given string.
359 # If possible, grab the one after the first heading
361 # It is possible to pass some options to determine how the stripping
362 # occurs. Currently supported options are
363 # strip:: Regex or String to strip at the beginning of the obtained
365 # min_spaces:: minimum number of spaces a paragraph should have
367 def Utils.ircify_first_html_par(xml_org, opts={})
368 if defined? ::Hpricot
369 Utils.ircify_first_html_par_wh(xml_org, opts)
371 Utils.ircify_first_html_par_woh(xml_org, opts)
375 # HTML first par grabber using hpricot
376 def Utils.ircify_first_html_par_wh(xml_org, opts={})
377 doc = Hpricot(xml_org)
379 # Strip styles and scripts
380 (doc/"style|script").remove
385 strip = Regexp.new(/^#{Regexp.escape(strip)}/) if strip.kind_of?(String)
387 min_spaces = opts[:min_spaces] || 8
388 min_spaces = 0 if min_spaces < 0
392 pre_h = pars = by_span = nil
395 debug "Minimum number of spaces: #{min_spaces}"
397 # Initial attempt: <p> that follows <h\d>
399 pre_h = Hpricot::Elements[]
401 doc.search("*") { |e|
407 pre_h << e if found_h
410 debug "Hx: found: #{pre_h.pretty_inspect}"
415 txt = p.to_html.ircify_html
416 txt.sub!(strip, '') if strip
417 debug "(Hx attempt) #{txt.inspect} has #{txt.count(" ")} spaces"
418 break unless txt.empty? or txt.count(" ") < min_spaces
421 return txt unless txt.empty? or txt.count(" ") < min_spaces
423 # Second natural attempt: just get any <p>
424 pars = doc/"p" if pars.nil?
425 debug "par: found: #{pars.pretty_inspect}"
428 txt = p.to_html.ircify_html
429 txt.sub!(strip, '') if strip
430 debug "(par attempt) #{txt.inspect} has #{txt.count(" ")} spaces"
431 break unless txt.empty? or txt.count(" ") < min_spaces
434 return txt unless txt.empty? or txt.count(" ") < min_spaces
436 # Nothing yet ... let's get drastic: we look for non-par elements too,
437 # but only for those that match something that we know is likely to
440 # Some blogging and forum platforms use spans or divs with a 'body' or
441 # 'message' or 'text' in their class to mark actual text. Since we want
442 # the class match to be partial and case insensitive, we collect
443 # the common elements that may have this class and then filter out those
444 # we don't need. If no divs or spans are found, we'll accept additional
445 # elements too (td, tr, tbody, table).
447 by_span = Hpricot::Elements[]
448 extra = Hpricot::Elements[]
449 doc.search("*") { |el|
450 next if el.bogusetag?
453 by_span.push el if el[:class] =~ AFTER_PAR_CLASS or el[:id] =~ AFTER_PAR_CLASS
455 extra.push el if el[:class] =~ AFTER_PAR_CLASS or el[:id] =~ AFTER_PAR_CLASS
458 if by_span.empty? and not extra.empty?
461 debug "other \#1: found: #{by_span.pretty_inspect}"
466 txt = p.to_html.ircify_html
467 txt.sub!(strip, '') if strip
468 debug "(other attempt \#1) #{txt.inspect} has #{txt.count(" ")} spaces"
469 break unless txt.empty? or txt.count(" ") < min_spaces
472 return txt unless txt.empty? or txt.count(" ") < min_spaces
474 # At worst, we can try stuff which is comprised between two <br>
477 debug "Last candidate #{txt.inspect} has #{txt.count(" ")} spaces"
478 return txt unless txt.count(" ") < min_spaces
479 break if min_spaces == 0
484 # HTML first par grabber without hpricot
485 def Utils.ircify_first_html_par_woh(xml_org, opts={})
486 xml = xml_org.gsub(/<!--.*?-->/m, '').gsub(/<script(?:\s+[^>]*)?>.*?<\/script>/im, "").gsub(/<style(?:\s+[^>]*)?>.*?<\/style>/im, "")
489 strip = Regexp.new(/^#{Regexp.escape(strip)}/) if strip.kind_of?(String)
491 min_spaces = opts[:min_spaces] || 8
492 min_spaces = 0 if min_spaces < 0
497 debug "Minimum number of spaces: #{min_spaces}"
498 header_found = xml.match(HX_REGEX)
501 while txt.empty? or txt.count(" ") < min_spaces
502 candidate = header_found[PAR_REGEX]
503 break unless candidate
504 txt = candidate.ircify_html
506 txt.sub!(strip, '') if strip
507 debug "(Hx attempt) #{txt.inspect} has #{txt.count(" ")} spaces"
511 return txt unless txt.empty? or txt.count(" ") < min_spaces
513 # If we haven't found a first par yet, try to get it from the whole
516 while txt.empty? or txt.count(" ") < min_spaces
517 candidate = header_found[PAR_REGEX]
518 break unless candidate
519 txt = candidate.ircify_html
521 txt.sub!(strip, '') if strip
522 debug "(par attempt) #{txt.inspect} has #{txt.count(" ")} spaces"
525 return txt unless txt.empty? or txt.count(" ") < min_spaces
527 # Nothing yet ... let's get drastic: we look for non-par elements too,
528 # but only for those that match something that we know is likely to
533 while txt.empty? or txt.count(" ") < min_spaces
534 candidate = header_found[AFTER_PAR1_REGEX]
535 break unless candidate
536 txt = candidate.ircify_html
538 txt.sub!(strip, '') if strip
539 debug "(other attempt \#1) #{txt.inspect} has #{txt.count(" ")} spaces"
542 return txt unless txt.empty? or txt.count(" ") < min_spaces
546 while txt.empty? or txt.count(" ") < min_spaces
547 candidate = header_found[AFTER_PAR2_REGEX]
548 break unless candidate
549 txt = candidate.ircify_html
551 txt.sub!(strip, '') if strip
552 debug "(other attempt \#2) #{txt.inspect} has #{txt.count(" ")} spaces"
555 debug "Last candidate #{txt.inspect} has #{txt.count(" ")} spaces"
556 return txt unless txt.count(" ") < min_spaces
557 break if min_spaces == 0
562 # This method extracts title, content (first par) and extra
563 # information from the given document _doc_.
565 # _doc_ can be an URI, a Net::HTTPResponse or a String.
567 # If _doc_ is a String, only title and content information
568 # are retrieved (if possible), using standard methods.
570 # If _doc_ is an URI or a Net::HTTPResponse, additional
571 # information is retrieved, and special title/summary
572 # extraction routines are used if possible.
574 def Utils.get_html_info(doc, opts={})
577 Utils.get_string_html_info(doc, opts)
578 when Net::HTTPResponse
579 Utils.get_resp_html_info(doc, opts)
582 @@bot.httputil.get_response(doc) { |resp|
583 ret.replace Utils.get_resp_html_info(resp, opts)
591 class ::UrlLinkError < RuntimeError
594 # This method extracts title, content (first par) and extra
595 # information from the given Net::HTTPResponse _resp_.
597 # Currently, the only accepted options (in _opts_) are
598 # uri_fragment:: the URI fragment of the original request
599 # full_body:: get the whole body instead of
600 # @@bot.config['http.info_bytes'] bytes only
602 # Returns a DataStream with the following keys:
603 # text:: the (partial) body
604 # title:: the title of the document (if any)
605 # content:: the first paragraph of the document (if any)
607 # the headers of the Net::HTTPResponse. The value is
608 # a Hash whose keys are lowercase forms of the HTTP
609 # header fields, and whose values are Arrays.
611 def Utils.get_resp_html_info(resp, opts={})
613 when Net::HTTPSuccess
614 loc = URI.parse(resp['x-rbot-location'] || resp['location']) rescue nil
615 if loc and loc.fragment and not loc.fragment.empty?
616 opts[:uri_fragment] ||= loc.fragment
618 ret = DataStream.new(opts.dup)
619 ret[:headers] = resp.to_hash
620 ret[:text] = partial = opts[:full_body] ? resp.body : resp.partial_body(@@bot.config['http.info_bytes'])
622 filtered = Utils.try_htmlinfo_filters(ret)
626 elsif resp['content-type'] =~ /^text\/|(?:x|ht)ml/
627 ret.merge!(Utils.get_string_html_info(partial, opts))
631 raise UrlLinkError, "getting link (#{resp.code} - #{resp.message})"
635 # This method runs an appropriately-crafted DataStream _ds_ through the
636 # filters in the :htmlinfo filter group, in order. If one of the filters
637 # returns non-nil, its results are merged in _ds_ and returned. Otherwise
640 # The input DataStream should have the downloaded HTML as primary key
641 # (:text) and possibly a :headers key holding the resonse headers.
643 def Utils.try_htmlinfo_filters(ds)
644 filters = @@bot.filter_names(:htmlinfo)
645 return nil if filters.empty?
647 # TODO filter priority
649 debug "testing htmlinfo filter #{n}"
650 cur = @@bot.filter(@@bot.global_filter_name(n, :htmlinfo), ds)
651 debug "returned #{cur.pretty_inspect}"
654 return ds.merge(cur) if cur
657 # HTML info filters often need to check if the webpage location
658 # of a passed DataStream _ds_ matches a given Regexp.
659 def Utils.check_location(ds, rx)
662 loc = [h['x-rbot-location'],h['location']].flatten.grep(rx)
666 return loc.empty? ? nil : loc
669 # This method extracts title and content (first par)
670 # from the given HTML or XML document _text_, using
671 # standard methods (String#ircify_html_title,
672 # Utils.ircify_first_html_par)
674 # Currently, the only accepted option (in _opts_) is
675 # uri_fragment:: the URI fragment of the original request
677 def Utils.get_string_html_info(text, opts={})
678 debug "getting string html info"
680 title = txt.ircify_html_title
682 if frag = opts[:uri_fragment] and not frag.empty?
683 fragreg = /<a\s+(?:[^>]+\s+)?(?:name|id)=["']?#{frag}["']?[^>]*>/im
686 if txt.match(fragreg)
687 # grab the post-match
693 c_opts[:strip] ||= title
694 content = Utils.ircify_first_html_par(txt, c_opts)
695 content = nil if content.empty?
696 return {:title => title, :content => content}
699 # Get the first pars of the first _count_ _urls_.
700 # The pages are downloaded using the bot httputil service.
701 # Returns an array of the first paragraphs fetched.
702 # If (optional) _opts_ :message is specified, those paragraphs are
703 # echoed as replies to the IRC message passed as _opts_ :message
705 def Utils.get_first_pars(urls, count, opts={})
709 while count > 0 and urls.length > 0
714 info = Utils.get_html_info(URI.parse(url), opts)
720 msg.reply "[#{idx}] #{par}", :overlong => :truncate if msg
724 debug "Unable to retrieve #{url}: #{$!}"
731 # Returns a comma separated list except for the last element
732 # which is joined in with specified conjunction
734 def Utils.comma_list(words, options={})
735 defaults = { :join_with => ", ", :join_last_with => _(" and ") }
736 opts = defaults.merge(options)
741 [words[0..-2].join(opts[:join_with]), words.last].join(opts[:join_last_with])
748 Irc::Utils.bot = Irc::Bot::Plugins.manager.bot